Abstract: The resource-constrained project scheduling problem with a flexible project network, random duration and rework is common in engineering practice. Under the complex conditions of random duration and random rework, it is necessary not only to solve the activity selection scheduling subproblem but also to solve the activity sequencing scheduling subproblem. The priority rule-based method is an efficient way to deal with complex stochastic resource-constrained project scheduling problems. Multiple classic priority rules for activity sequencing and activity selection were summarized, and test instances covering different problem characteristics were constructed based on benchmark instances. The performance of multiple rules in different problem environments was compared through extensive simulation experiments. The results indicate that paired-rules are significantly better than single rules, and the best performing paired-rule differs from the conclusions of existing literature. In addition, it was found that project flexibility and resource strength have a significant impact on the performance of rules, while the influence of other factors is relatively small. The conclusion drawn has important reference value for selecting appropriate project scheduling rules in engineering practice.